Automated Computational Design of Composite Li-ion Battery Electrodes Microstructures

PI：Soghrati, Soheil 提案编号：1608058 拟议研究的目标是开发一种集成计算方法来建模和预测锂离子电池中电极的性能。了解和开发锂离子电池模型的价值非常重要，因为它们非常常用于能量存储。拟议的研究可以通过取代目前应用的电极设计试错方法来提高混合动力和电动汽车应用的性能。这项研究将通过集成计算材料工程（ICME）方法量化锂离子电池（LIB）中使用的复合电极的最佳微观结构特征。虽然锂离子电池是二次储能的主要技术，但仍需要取得重大进展来提高其在交通运输领域的应用的功率和能量密度。通过优化电极设计以最大化电导率和表面积，可以显着提高锂离子电池的性能。然而，由于无法真实地模拟这种复合材料的复杂异质结构，目前设计过程主要以试错实践为主，通常会导致设计次优，并导致大量的开发时间和成本。该提案旨在通过开发新的设计优化框架来克服这一障碍，该框架包括：（i）提取电极微结构的分层多尺度成像数据； (ii) 开发基于成像数据自动创建电极异质结构的真实虚拟模型的能力； (iii) 使用先进的有限元方法模拟电极的多物理场响应； (iv) 通过纽扣电池原型的实验测试来验证模型； (v) 通过多目标遗传算法确定产生最高功率和能量密度的最佳微观结构。分层界面富集有限元方法（HIFEM）将作为主要计算引擎，用于模拟充电/放电循环期间电极的多物理场行为。 HIFEM 使用完全独立于问题形态的简单结构化网格，产生与标准 FEM 相同的精度和收敛速度。 HIFEM 将与依赖非均匀有理基样条 (NURBS) 的虚拟原型算法集成，以基于涉及 X 射线显微断层扫描、聚焦离子束断层扫描和电子显微镜的分层成像数据创建复合电极的真实 3D 微观结构模型。为了准确预测与各种电极设计相关的功率和能量密度，将采用高保真度计算模型来模拟锂离子电池的电化学机械响应。此外，还将为该自动化计算管道部署完全并行的计算模块，以创建电极的虚拟微观结构模型并模拟其多物理场行为。如果成功，拟议的项目将导致开发用于复合电极虚拟设计的计算框架。通过这项研究获得的基础知识将使在其产品中大量使用锂离子电池的多个行业受益，例如汽车、航空航天和便携式电子产品。此外，在该项目期间开发的建模功能可用于处理具有类似微观结构复杂性的更广泛的 ICME 问题。为了将研究、推广和教育整合到该项目中，将执行以下任务： (i) 强大的 www 和社交媒体影响力，以促进向公众和科学界的推广； (ii) 参加俄亥俄州立大学 (OSU) 组织的 K-12 外展计划； (iii) 培训和指导研究生和本科生，并将研究成果纳入课程； (iv) 通过将工程研究转化为 K-8 计划向中学生进行推广，该计划还将聘请俄勒冈州立大学的本科生研究助理作为职业大使。

PI: Soghrati, SoheilProposal Number: 1608058The goal of the proposed research is to develop an integrated computational approach to model and predict the performance of electrodes in Li-ion batteries. The value of understanding and developing models for Li-ion batteries is very significant, since these are very commonly used for energy storage. The proposed research could lead to improved performance for applications in hybrid and electric vehicles, by replacing the currently applied trial and error approach to electrode design. This research will quantify the optimal microstructural features of composite electrodes used in lithium-ion batteries (LIBs) via an integrated computational materials engineering (ICME) approach. While LIBs are the primary technology for secondary energy storage, significant advancements are still needed to improve their power and energy densities for application in the transportation sector. Remarkable improvement in the LIBs performance can be achieved by optimizing the electrode design to maximize the conductivity and surface areas. However, due to inability to realistically model the intricate heterostructure of this composite material, design processes are currently dominated by trial-and-error practices, often leading to sub-optimal designs and significant development time and cost. This proposal aims at overcoming this barrier by developing a new design optimization framework consisting of: (i) extracting hierarchical multiscale imaging data of the electrode microstructure; (ii) developing the ability to automatically create realistic virtual models of the electrode heterostructure based on imaging data; (iii) simulating the multiphysics response of the electrode using an advanced finite element method; (iv) validating the models via experimental testing of coin-cell prototypes; (v) identifying the optimal microstructures that yield highest power and energy densities via a multi-objective genetic algorithm. A hierarchical interface-enriched finite element method (HIFEM) will serve as the main computational engine for simulating the multiphysics behavior of the electrode during charge/discharge cycles. The HIFEM yields the same precision and convergence rate as those of the standard FEM using simple structured meshes that are completely independent of the problem morphology. The HIFEM will be integrated with a virtual prototyping algorithm relying on Non-Uniform Rational Basis Splines (NURBS) to create realistic 3D microstructural models of the composite electrode based on hierarchical imaging data involving x-ray microtomography, focused ion beam tomography, and electron microscopy. To accurately predict the power and energy densities associated with various designs of the electrode, a high fidelity computational model will be implemented to simulate the electro-chemo-mechanical response of the LIB. Furthermore, a fully parallel computing module will be deployed for this automated computational pipeline to both create virtual microstructural models of the electrode and simulate its multiphysics behavior. If successful, the proposed project will lead to the development of a computational framework for the virtual design of composite electrodes. The fundamental knowledge generated through this research will benefit several industries heavily using LIBs in their products, such as the automotive, aerospace, and portable electronics. Moreover, the modeling capabilities that will be developed during the term of this project can be employed for the treatment of a broader range of ICME problems with similar microstructural complexities. To integrate the research, outreach, and education in this project, these tasks will be pursued: (i) strong www and social media presence to facilitate outreach to the public and scientific communities; (ii) participating in K-12 outreach programs organized by Ohio State University (OSU); (iii) training and mentoring of graduate and undergraduate students and integrating research outcomes into the curriculum; (iv) outreach to middle school students via Translating Engineering Research to K-8 program, which will also engage OSU's undergraduate research assistants as career ambassadors.